Cryptonite: A Secure and Performant Data Repository on

Public Clouds
Abstract:
Cloud storage has become immensely popular for maintaining synchronized copies of files and
for sharing documents with collaborators. However, there is heightened concern about the
security and privacy of Cloud-hosted data due to the shared infrastructure model and an implicit
trust in the service providers. Emerging needs of secure data storage and sharing for domains like
Smart ower !rids, which deal with sensitive consumer data, re"uire the persistence and
availability of Cloud storage but with client-controlled security and encryption, low key
management overhead, and minimal performance costs. Cryptonite is a secure Cloud storage
repository that addresses these re"uirements using a StrongBox model for shared key
management. #e describe the Cryptonite service and desktop client, discuss performance
optimizations, and provide an empirical analysis of the improvements.
EXISTING SSTE!:
E$isting Cloud storage services only provide basic access control
mechanisms, and the limited research on secure, shared Cloud repositories often re"uire
e$tensive deployment of infrastructure services that undermines their manageability. Specifically,
there are two classes of current solutions. %ne uses simple file en&decryption on the client side,
managed through shared keys, such that only encrypted files are hosted in Clouds.
Disadvantage:
'ownloading time for large data will be high.
erformance for the 'ownloading time is high.
Compare to #indows azure platform it has some failures.
PR"P"SED SSTE!:
(here are two classes of current solutions. %ne uses simple file
en&decryption on the client side, managed through shared keys, such that only encrypted files are
hosted in Clouds. #hile this works for a small number of files and users, key sharing and
management becomes unsustainable as the number of users sharing data grows. ) second
solution is to allow the Cloud providers or a third-party offer more advanced security services
and manage user credentials such as dropbo$.
Advantage:
(he file downloading performance is high.
*arge files can be downloaded easily.
(he encrypted +iles are downloaded slowly compared with normal downloaded.
(he different speeds of downloading from server are compared.
!"D#$ES:
%& !odule
(here are two classes of current solutions. %ne uses simple file en&decryption on
the client side, managed through shared keys, such that only encrypted files are hosted in Clouds.
#hile this works for a small number of files and users, key sharing and management becomes
unsustainable as the number of users sharing data grows. ) second solution is to allow the Cloud
providers or a third-party offer more advanced security services and manage user credentials
such as dropbo$.
'& !odule
(he concept of a StrongBox file enables scalable key management by securing multiple files that
share the same permissions using ,ust the single global public-private keypair for each user.
-ntuitively, a StrongBox represents a uni"ue combination of permission .such an access control
list/, with a specified owner, list of writers and list of readers.
(& !odule
(he Cryptonite Client Library (CCL) is responsible for performing
cryptographic operations on the client-side on behalf of the user, and interacting with the
Cryptonite service. #hile the CC* is provided for convenience, users or their trusted developers
can implement this library themselves using standard cryptographic algorithms and web service
protocols for absolute trust when handling the plainte$t files and the private keys of the users.
CC* performs encryption, signing and header generation to create the '+iles from plainte$t files
using the private keys available for the user and shared public keys available for other users
using the 0- infrastructure.
)& !odule
(he final optimization we perform is in e$tending data parallelism to other stages in the
flow, in particular, for block encryption .E/ on the client side. #e also decouple the validation
.1/ from the block upload .2/ on the serviceside.
we present a detailed empirical evaluation of the performance characteristics of the Cryptonite
service and CC*, and the impact of different optimization strategies we have proposed. #e use
3effective bandwidth4 as our metric, defined as the lainte$t +ile Size
Endto End Operation Time Mbits, since this captures the tangible impact for the user.
System Re*uirements:
+ard,are Re*uirements:
System 5 entium -1 6.7 !Hz.
Hard 'isk 5 78 !9.
+loppy 'rive 5 :.77 ;b.
;onitor 5 :< 1!) Colour.
;ouse 5 *ogitech.
=am 5 <:6 ;b.
Soft,are Re*uirements:
%perating system 5 #indows >&8?.
Server 5 #ampServer&>ampServer.
Coding *anguage 5 H
'ata 9ase 5 ;y-S@*